Synthesis and evaluation of hetero- and homo-dimers of ribosome-targeting antibiotics: Antimicrobial activity, in vitro inhibition of translation, and drug resistance

Abstract

In this study, we describe the synthesis of a full set of homo- and hetero-dimers of three intact structures of different ribosome-targeting antibiotics: tobramycin, clindamycin, and chloramphenicol. Several aspects of the biological activity of the dimeric structures were evaluated including antimicrobial activity, inhibition of in vitro bacterial protein translation, and the effect of dimerization on the action of several bacterial resistance mechanisms that deactivate tobramycin and chloramphenicol. This study demonstrates that covalently linking two identical or different ribosome-targeting antibiotics may lead to (i) a broader spectrum of antimicrobial activity, (ii) improved inhibition of bacterial translation properties compared to that of the parent antibiotics, and (iii) reduction in the efficacy of some drug-modifying enzymes that confer high levels of resistance to the parent antibiotics from which the dimers were derived.

INTRODUCTION

Protein biosynthesis is one of the fundamental processes required for all living cells, and the prokaryotic ribosome is a target of a large number of clinically useful antibiotics.^1–4 Bacterial ribosome-targeting antibiotics inhibit protein synthesis by interfering with the process of messenger RNA translation or by preventing the formation of peptide bond of the nascent protein. These effects result from the binding of these antibiotics to the aminoacyl-tRNA binding domain (A-site),^5–7 to the peptidyltransferase domain (P-site),^8–10 or to elements on both sites.¹⁰

Prolonged and inappropriate use of the various bacterial ribosome-targeting antibiotics enhanced the evolution of several bacterial resistance mechanisms that can be divided into four types: (i) inactivation through enzymatic modifications,¹¹ (ii) reduction in intracellular concentration through efflux pump proteins,^{12, 13} (iii) reduction in membrane permeability,¹⁴ and (iv) structural modifications in the ribosomal target sites.¹⁵ Amongst ribosomal target modifications that lead to antibiotic resistance are mutations in rRNA, mutations in ribosomal proteins, and rRNA methylation catalyzed by several rRNA methyltransferases.^16–18

The development of covalently linked dimers composed of intact clinically used antibiotics that inhibit similar or different bacterial targets has been widely explored.^19–21 The potential benefits of this strategy include the possibility of an expanded spectrum of activity, reduced potential for development of bacterial resistance, and an increase in specificity for bacterial cells that leads to reduced toxicity to the host. Several major obstacles stand in the way of developing novel antimicrobials by dimerization of existing antimicrobial agents. These obstacles include cell permeability problems, unexpected binding to additional cellular targets, reduction in target affinity due to the chemical modification of the parent antibiotics, and unexpected bacterial resistance.

A different approach for the design of novel ribosome-targeting antimicrobial agents is based on the design of hybrid antibacterials derived from fragments of ribosome targeting antibiotics.^{22, 23} This should result in antibacterials that bind to rRNA with higher affinity, evade known modes of resistance, and more effectively cause bacterial cell death. Most bacterial ribosome-targeting antibiotics have a single binding site on the ribosome; therefore only one monomer unit of a designed antimicrobial homo-dimer can bind to the target at one time. However, homo-dimerization increases the local concentration of the drug and may therefore lead to improvement in target inhibition. Moreover, if an antimicrobial agent has more than one possible binding domain and these domains are in close proximity on the ribosome, homo-dimerization may result in improved target affinity. One such example was previously reported. Using surface plasmon resonance (SPR) experiments, it was demonstrated that the pseudo-disaccharide neamine (NEA, Figure 1), a segment of several common aminoglycoside (AG) antibiotics, bound to an rRNA sequence that is a model for the prokaryotic A-site in a 2:1 ratio, suggesting that AGs can bind to two sites that are in close proximity.²⁴ Based on these observations, NEA homo-dimers were designed and synthesized using various linkers. These homo-dimers had significantly increased affinity for the target A-site rRNA compared to the parent monomer and bound to the target in a 1:1 ratio. Furthermore, some NEA homo-dimers were poor substrates for AG-modifying enzymes (AMEs) that confer resistance to AGs through chemical modifications and demonstrated more potent antimicrobial activity than the parent monomer NEA. Finally, homo- and hetero-dimers^25–27 and oligomers²⁸ of known rRNA-targeting antibiotics have been explored for targeting viral RNAs or endogenous RNAs involved in disease in an attempt to inhibit their biological activity.

Figure 1.

Structures of neamine, tobramycin (1), clindamycin (2), and chloramphenicol (3).

We are particularly interested in systematically exploring the potential that lies in the development of novel antimicrobial agents through the dimerization of intact structures of ribosome-targeting antibiotics. For that purpose, we designed and synthesized a complete combinatorial set of homo- and hetero-dimers composed of three different ribosome-targeting antibiotics and studied several aspects of their antimicrobial activity. The three antibiotics that served as starting materials were tobramycin (1), clindamycin (2), and chloramphenicol (3). Each chosen antibiotic binds to a different site on the bacterial ribosome (Figure 1).

The AG 1, which targets the A-site, is particularly effective for the treatment of infections caused by the pathogenic Gram-negative bacterium Pseudomonas aeruginosa.²⁹ Several mechanisms of resistance have evolved against 1: reduction in the intracellular concentration of this antibiotic by efflux pump proteins or through reduced membrane permeability, structural modifications of the 16S rRNA leading to reduced target affinity, and deactivation by AMEs. AMEs are divided into three families: AG nucleotidyltransferases (ANTs), AG phosphotransferases (APHs), and AG acetyltransferases (AACs).³⁰ In addition to the increasing problem of antibiotic resistance, like other AGs, 1 causes reversible nephrotoxic and irreversible ototoxic side effects.³⁰

A 2.54-Å resolution structure of a complex of an A-site 16S rRNA sequence and 1 revealed that the five amine groups of 1, which are bound in their ammonium form, participate in hydrogen bonds with either bridging water molecules or directly with the target 16S rRNA nucleotides.³¹ All of the alcohols of this AG, with the exception of its single C-6" primary alcohol, also interact with A-site nucleotides. We therefore chose to chemically modify the 6" primary alcohol of 1 to facilitate the attachment of the linker unit that was used for the preparation of the desired tobramycin-derived dimers.

The lincosamide antibiotic 2 is effective against staphylococci and streptococci infections.³² Bacterial resistance to 2 occurs rarely through enzymatic deactivation and more prevalently as a result of 23S rRNA modifications, which reduce affinity of 2 for the ribosome and decrease its clinical efficacy.^{33, 34} The structure of 2 bound to the 50S subunit of the E. coli ribosome indicated that the C-7 chloride atom does not specifically interact with the ribosomal RNA or proteins.⁸ We therefore chose the C-7 of 2 as a suitable position for the attachment of a linker unit through nucleophilic displacement of the chloride atom leading to inversion of the absolute configuration of the C-7 position of 2.

Finally, the antibiotic 3 is widely used against a broad spectrum of Gram-positive and Gram-negative bacteria including anaerobes.³⁵ Resistance to this antimicrobial agent occurs through reduced membrane permeability, mutation of the 50S ribosomal subunit, and deactivation by three different types of enzymes, chloramphenicol acetyltransferase (CAT), chloramphenicol phosphotransferase (CPT), and chloramphenicol nitroreductase (CNR).³⁶ Treatment with 3 may cause severe side effects such as aplastic anemia and bone marrow suppression, and there is a need for analogs of 3 that are safer for clinical use.³⁷ The antibiotic 3 binds specifically to nucleotides of the 23S rRNA and prevents peptide bond formation.^{8, 10} X-ray crystallography of 3 bound to the 50S subunit of the E. coli ribosome revealed that the two chloride atoms of its dichloroacetamide segment appear to have no significant role in target binding. Therefore, we chose to modify this position of 3 for the generation of a linker attached monomer unit to be used for the generation of the desired dimeric structures.

RESULTS AND DISCUSSION

Synthesis of ribosome-targeting antibiotic dimers

Following the robust strategy for the synthesis of the homo- and hetero-dimers of AGs that was reported by Tor and co-workers,²⁷ we chose to attach a bis(2-mercaptoethyl) ether unit to each of the chosen ribosome-targeting antibiotics (Scheme 1). To prepare derivatives of 1, compound 4 was reacted with bis(2-mercaptoethyl) ether to form compound 5 (Scheme 1A) as was previously reported.²⁷ Boc groups were removed by treatment with neat TFA to yield derivative 6. Compound 5 was then reacted with 2,2'-dipyridyldisulfide to form the activated disulfide derivative 7 that was used for the preparation of tobramycin-based disulfide-linked homo- and hetero-dimers.

Scheme 1.

Synthetic schemes for the preparation of A. Tobramycin monomer building blocks, B. Clindamycin monomer building block and of clindamycin-clindamycin dimer 10, and C. chloramphenicol monomer building blocks.

The C-7 chloride of 2 served as a leaving group and was displaced by a bis(2-mercaptoethyl) ether unit to form the intermediate S_N2 product, which was per-acetylated to facilitate the isolation of the clindamycin analog 8 (Scheme 1B). Two different products were obtained depending on conditions used for the removal of acetyl groups of 8. When 0.04 M of compound 8 was treated with an excess of sodium methoxide in MeOH and stirred at ambient temperature for 3 hours, the deprotected monomer 9 was obtained as the major product (83% isolated yield). However, when the reaction was ten-fold more concentrated (0.4 M of compound 8), and 8 was treated overnight with an excess of sodium methoxide at ambient temperature, the disulfide linked clindamycin-derived homo-dimer 10 was obtained as the major product (75% isolated yield).

Compound 14 was obtained by preparing 12 from 2-amino-1-(4-nitrophenyl)-1,3-propanediol.³⁸ The α-chloride atom of 12 was displaced by bis(2-mercaptoethyl) ether to yield 13. Reaction of 13 with 2,2'-dipyridyldisulfide yielded the activated disulfide 14 (Scheme 1C).

With the building block monomers in hand, we then prepared the rest of the five combinations of homo- and hetero-dimers (16, 18, 22, 24 and 26, Scheme 2) to form the complete set of six possible combinations of disulfide-linked dimers of the three chosen ribosome-targeting antibiotics. In the case of the tobramycin-containing dimers, all of the Boc protecting groups were readily removed by short treatment with neat TFA. The synthesis of homo-dimer 16 was previously reported; this dimeric AG was a significantly better inhibitor of the activity of the Tetrahymena ribozyme than the parent antibiotic 1.²⁷

Scheme 2.

Synthetic scheme for the preparation of homo- and hetero-dimers derived of 1, 2, and 3.

To test the effect of linker length, we also prepared three dithioether-based dimers with short linkers. Nucleophilic displacement of the α-chloride atom of 12 by thiol 5 gave the hetero-dimer 19, which was converted to 20, the short linker analog of hetero-dimer 22, by removal of the Boc groups. A similar strategy was used for synthesis of hetero-dimer 23 and for the preparation of homo-dimer 25, the short linker analogs of hetero-dimers 24 and 26, respectively.

Correlation between antimicrobial activity and in vitro inhibition of prokaryotic translation

We tested the antimicrobial potency of the synthetic antibiotic dimers against a selection of Gram-positive and Gram-negative pathogens (Table 1), and analyzed the relationship between their antimicrobial activities and protein translation inhibition properties by performing in vitro translation assays using extracts containing E. coli ribosomes (Table 2).

Table 1.

Antibacterial activity of homo- and hetero-dimers and parent AGs against Gram-positive (+) and Gram-negative (−) bacterial strains^a: MIC values (µg/mL).

Cpd #	A (+)	B (+)	C (+)	D (+)	E (+)	F (+)	G (+)	H (−)	I (+)	J (−)	K (+)	L (+)
Parent drugs
1	2.3	9.4	0.6	75	18.8	37.5	9.4	9.4	9.4	18.8	0.3	64
2	0.6	>150	9.4	1.2	4.7	0.6	<0.3	2.3	<0.3	4.7	1	1
3	2.3	1.2	18.8	4.7	9.4	9.4	4.7	<0.3	4.7	0.6	8	4
Parent compounds with linker
6	2.3	75	2.3	150	18.8	75	18.8	18.8	37.5	37.5	16	128
9	9.4	>150	75	9.4	37.5	4.7	1.2	37.5	2.3	75	4	2
13	>150	>150	>150	150	>150	>150	>150	37.5	150	>150	>128	128
Homo-dimers
10	2.3	>150	18.8	2.3	4.7	0.6	0.6	9.4	2.3	37.5	1	>128
16	0.6	37.5	2.3	75	37.5	37.5	9.4	18.8	37.5	9.4	16	64
25	>150	>150	>150	>150	>150	>150	>150	>150	>150	>150	>64	>128
26	150	>150	>150	75	9.4	0.6	>150	18.8	75	150	>64	>128
Hetero-dimers
18	1.2	150	4.7	9.4	18.8	2.3	4.7	4.7	2.3	75	2	64
20	150	>150	18.8	>150	>150	>150	>150	>150	>150	150	8	8
22	4.7	150	4.7	150	18.8	150	9.4	4.7	75	37.5	16	16
23	18.8	>150	>150	18.8	75	18.8	18.8	9.4	2.3	75	4	>128
24	2.3	>150	18.8	2.3	4.7	0.6	<0.3	4.7	<0.3	150	2	>128

^aB. subtilis 168 (A), B. subtilis 168 with AAC(6')/APH(2")-pRB374 (B), M. smegmatis str. MC2 155 (C), B. cereus ATCC 11778 (D), L. monocytogenes ATCC 19115 (E), S. aureus ATCC 29213 (F), S. aureus NorA (G) E. coli MC1061 (H), B. anthracis 34F2 Sterne strain (I), H. influenzae ATCC 51907 (J), S. epidermidis ATCC 12228 (biofilm negative) (K), and S. pyogenes serotype M12 str. MGAS9429 (L).

Table 2.

Activities of homo- and hetero-dimers and parent AGs in an in vitro prokaryotic translation assay.

Compound #	IC50 (µM)
Parent drugs
1	0.015 ± 0.001
2	18.1 ± 0.8
3	6.4 ± 0.4
Parent compounds with linker
6	0.43 ± 0.01
9	>100
13	48.7 ± 1.8
Homo-dimers
10	6.1 ± 1.1
16	0.27 ± 0.01
25	32.9 ± 2.6
26	25.6 ± 2.1
Hetero-dimers
18	0.98 ± 0.05
20	1.49 ± 0.05
22	1.00 ± 0.16
23	14.7 ± 1.2
24	18.3 ± 0.3

Compared to the parent antibiotic 1 (IC₅₀ = 0.015 ± 0.001 µM), compound 6 was approximately 29-fold less potent as an inhibitor of in vitro translation (IC₅₀ = 0.43 ± 0.01 µM); the corresponding homo-dimer 16 (IC₅₀ = 0.27 ± 0.01 µM) was 18-fold less potent than 1. Antimicrobial activities were strain dependent: In the case of B. subtilis 168 (strain A) and of H. influenzae ATCC 51907 (strain J), homo-dimer 16 was one to two double dilutions more potent than parent antibiotic 1 and two double dilutions more potent than the monomer building block 6 (Table 1). Monomer 6 and homo-dimer 16 were less potent than 1 against bacterial strains B, C, H, I, and K (Table 1). In general, the MIC values of compound 6 and of the homo-dimer 16 were similar in the majority of the tested strains. The fact that both compound 6 and homo-dimer 16 were poor inhibitors of in vitro translation compared to 1 but were potent against several of the tested bacterial strains suggests that the mode of action of these compounds may differ from that of the parent AG 1.

Compared to the parent drug 2, compound 9 proved a poor inhibitor of prokaryotic translation with IC₅₀ values higher than 100 µM, at least 5-fold higher than that of the parent antibiotic 2. In contrast to the monomer building block 9, the homo-dimer 10 was about 3-fold more potent in the in vitro translation assay than the parent antibiotic 2 (IC₅₀ values of 6.1 ± 1.1 µM and 18.1 ± 0.8 µM, respectively). Although compound 9 was a poor inhibitor of in vitro translation, it had MIC values of 1.2–9.4 µg/mL against seven of the tested strains, suggesting that the activity of this compound is not due to inhibition of prokaryotic translation. The dimer of 9, homo-dimer 10, was a considerably better inhibitor of in vitro translation than 2. Homo-dimer 10 exhibited good antimicrobial activity against eight of the tested bacterial strains with MICs of 0.6–9.4 µg/mL. Moreover, compound 10 was significantly more active than its monomer building block 9 and comparable in activity to the parent antibiotic 2 against seven of the tested bacterial strains (B–G, K).

In the in vitro prokaryotic translation assay, the chloramphenicol-based monomer 13 was close to 8-fold less potent than the parent antibiotic 3 (Table 2). Compared to compound 13, the chloramphenicol-based homo-dimers 25 and 26 exhibited a modest improvement in inhibition of prokaryotic translation (IC₅₀ values 48.7 ± 1.8 µM for 13, 32.9 ± 2.6 µM for 25, and 25.6 ± 2.1 µM for 26). Both the compounds 13 and 25 demonstrated poor antimicrobial activities compared to 3 against all of the tested bacterial strains. On the other hand, homo-dimer 26 was as potent as the parent against L. monocytogenes ATCC 19115 (strain E, MIC of 9.4 µg/mL) and was approximately 16-fold more potent than 3 against S. aureus ATCC 29213 (strain F).

Interestingly, the measured IC₅₀ values of the inhibition of the in vitro prokaryotic translation of all hetero-dimers (18, 20, 22, 23, and 24) were in the range bracketed by activities of the parent antibiotics from which these hetero-dimers were derived from (Table 2). For example, the IC₅₀ value of the tobramycin-clindamycin hetero-dimer 18 in the in vitro translation assay was 0.98 ± 0.05 µM, which was approximately 65-fold higher than that of 1 and about 18-fold lower than that of 2. The linker type and length had almost no effect on the inhibition of prokaryotic translation (Table 2, compare 25 and 26, 20 and 22, as well as 23 and 24). On the other hand, the linker type and length had a significant effect on the antimicrobial activity: The short linker homo- and hetero-dimers (20, 23, and 25) were far less potent antimicrobial agents than the corresponding long disulfide linker hetero-dimers (22, 24, and 26) against all of the tested strains (Table 1). The only exceptions were observed in the case of compounds 20 and 22; the short linker hetero-dimer 20 was one double dilution more potent than the long disulfide linker hetero-dimer 22 against S. epidermidis ATCC 12228 (biofilm negative, strain K) and against S. pyogenes serotype M12 str. MGAS9429 (strain L).

To investigate the potential effect of active efflux on our compounds, we determined MIC values against S. aureus that expresses NorA (strain G), a protein that mediates the active efflux of hydrophilic drugs from the cells to confer resistance. We observed that this mechanism of resistance does not appear to affect molecules containing 1- or 2-based building blocks as the MIC values of compounds 10, 16, and 18 were comparable to those of the parent drugs.

Effect of dimers on drug-modifying enzyme activity

To investigate the effect of covalently linking two intact ribosome-targeting antibiotics on the activity of drug-modifying enzymes responsible for the deactivation of the parent drugs in a wide spectrum of antibiotic-resistant bacterial strains, we selected six AMEs and three chloramphenicol-modifying enzymes against which we tested all tobramycin- and chloramphenicol-derived homo- and hetero-dimers (Figure 2).

Figure 2.

A. Activities of various AMEs on tobramycin-containing molecules (6, 16, 18, 20, and 22) relative to 1. B. Activities of various chloramphenicol-modifying enzymes on chloramphenicol -containing molecules (13, 20, 22–26) relative to 3.

The substrate preference of the nine tested drug-deactivating enzymes varied. Of the six tested tobramycin-targeting AMEs, the acetyltransferase activity of the bifunctional AAC(6')/APH(2"), the acetyltransferase activity of AAC(6')-Ib', and the nucelotidyltransferase activity of ANT(4') were significantly less against tobramycin-derived dimers 16, 18, 22, and 20 than against the parent AG 1 (Figure 2A and Table S1). The activity of all AMEs was set to 100% for the parent antibiotic 1 for comparison purposes. Dimers 16, 18, 22 and 20 were poor substrates of AAC(6')-Ib' (39%, 20%, 31%, and 37%, respectively). The tobramycin derivatives 16, 18, and 22 were also poorer substrates of AAC(6')/APH(2") (34%, 31%, and 34%, respectively). Moreover, tobramycin-derived dimers 18 and 22 were poorer substrates of ANT(4') (38% and 28%, respectively) than 16 and 20 (90% and 63%, respectively). Tobramycin derivative 6, with the thiol-containing linker, could only be tested against ANT(4') as our assay for AAC enzymes relies on detection of free thiol. Compound 6 was a poorer substrate of ANT(4') (49%) than the parent AG 1 (100%). The opposite effect was observed for the other three AMEs tested; compared to the parent AG 1, all of the tobramycin-based dimers (16, 18, 20 and 22) were superior substrates for the acetyltransferases AAC(3)-IV, AAC(2')-Ic, and Eis (Figure 2A). Overall, compounds 18 and 22 behaved similarly with all AMEs and were poorer substrates then 16 and 20. With AAC(3)-IV, AAC(2')-Ic, and Eis, compounds 18 and 22 were 1.4-, 1.6-, and ~2.6-fold better substrates, respectively than 1, whereas compound 16 was a 2-, 2.4-, and 4.1-fold better substrate than 1, and compound 20 was 1.8-, 2.0-, and 2.9-fold better than 1.

The chloramphenicol-derived homo-dimer 25 was a poor substrate for all of the three chloramphenicol-modifying enzymes: CPT, CNR, and CAT_I (Figure 2B and Table S1). However, this homo-dimer demonstrated poor antimicrobial activity against all bacterial strains tested (Table 1). With the exception of homo-dimer 25, CNR readily modified all chloramphenicol-derived dimers (20, 76%; 22, 117%; 23, 133%; 24, 174%; and 26, 220%) at least as well as it modified the parent antibiotic 3 (100%) (Figure 2B). The chloramphenicol acetyltransferase CAT_I is commonly found in chloramphenicol-resistant bacteria. Of the tested chloramphenicol-modifying enzymes, all of the chloramphenicol-derived dimers (20, 82%; 22, 70%; 23, 10%; 24, 69%; and 26, 73%) were poorer substrates of CAT_I than the parent antibiotic 3 (100%). With the CPT, all chloramphenicol-derived dimers behaved similarly to 3, with relative activities ranging from 91–139%. With the exception of dimers 25 and 26, the length and type of the linker did not significantly affect the resistance profile of the dimers. The dimer approach proved to be more effective in interfering with the drug-modifying activities of the tested AMEs than with the activities of the three tested chloramphenicol-modifying enzymes.

CONCLUSION

In this study we describe the synthesis of nine homo and hetero-dimers composed of combinations of the three intact structures of the bacterial ribosome-targeting antibiotics 1–3. The effect of the homo- and hetero-dimerization approach was evaluated by testing the antimicrobial activity, the potency of the dimers as inhibitors of in vitro prokaryotic translation, and the ability of nine different drug-deactivating enzymes to modify the dimers derived from chloramphenicol and/or tobramycin.

Of the nine dimeric structures that were evaluated in this study, two homo-dimers exhibited potent antibacterial activities, which were in some cases superior to those of the parent antibiotics from which they were derived. Both the tobramycin-derived homo-dimer 16 and the clindamycin-derived homo-dimer 10 were comparable and in some cases more potent antimicrobials than the parents 1 and 2 against some of the twelve bacterial strains in our tested panel. The results of the in vitro inhibition of prokaryotic translation assay suggest that 1 and the tobramycin-derived homo-dimer 16 may act through different modes of action. On the other hand, the clindamycin-derived homo-dimer 10 was approximately 3-fold more potent than the parent 2 as an inhibitor of prokaryotic translation. These results demonstrate that homo-dimerization of intact ribosome targeting antibiotics may lead to a change in the mode of action or to improved target inhibition.

Of the five hetero-dimers, compound 18 demonstrated a broader antimicrobial activity spectrum than either of the parent antibiotics. As was the case for several of the hetero-dimers, the in vitro inhibition of prokaryotic translation potency of compound 18 was between that of its parent antibiotics 1 and 2.

The effect of drug-modifying enzymes on the dimers depended on the enzyme. Some drug-deactivating enzymes modified dimers more efficiently than they did the parent antibiotics. On the other hand, the two potent tobramycin-derived compounds 16 and 18 were approximately 2.5–5 times less prone to deactivation by the two AG-modifying enzymes AAC(6')/APH(2") and AAC(6')-Ib' than was the parent AG 1. AAC(6') enzymes account for high levels of AG resistance and are prevalently found amongst AG-resistant bacterial strains. Hence both homo- and hetero-dimerization of ribosome-targeting antibiotics can be used to combat certain AG resistance mechanisms. This study demonstrates that even though the diversity amongst bacterial pathogens makes it impossible to predict all of the possible effects of drug dimerization, dimerization is a promising strategy for development of novel inhibitors of bacterial protein translation. Further exploration of this strategy should result in novel antimicrobial agents that may bind to the prokaryotic ribosome through unique sets of interactions that differ from those of the parent antibiotics.

EXPERIMENTAL SECTION

Bacterial strains, plasmids, materials, and instrumentation

The bacterial strains tested in this study were obtained from a variety of sources. Bacillus subtilis 168 (A) was obtained from the Bacillus Genetic Stock Center (Columbus, OH, USA). Bacillus subtilis 168 with AAC(6')/APH(2")-pRB374 (B) was prepared as we previously described.³⁹ Mycobacterium smegmatis str. MC2 155 (C) was a generous gift from Dr. Sabine Ehrt (Weill Cornell Medical College). Bacillus cereus ATCC 11778 (D), Listeria monocytogenes ATCC 19115 (E), Staphylococcus aureus ATCC 29213 (F), and Escherichia coli MC1061 (H) were provided by Prof. Paul J. Hergenrother (University of Illinois at Urbana-Champaign). Staphylococcus aureus NorA (G) was a gift from Prof. David Sherman (University of Michigan). Bacillus anthracis 34F2 Sterne strain (I) was a gift from Prof. Philip C. Hanna (University of Michigan). Haemophilus influenzae ATCC 51907 (J) and Staphylococcus epidermidis ATCC 12228 (biofilm negative) (K) were purchased from ATCC. Streptococcus pyogenes M12 str. MGAS9429 (L) was provided by Prof. Doron Steinberg (The Hebrew University of Jerusalem). Chemically competent E. coli TOP10 and BL21 (DE3) used in cloning were purchased from Invitrogen. The Int-pET19b–pps plasmid was a gift from Dr. Tapan Biswas (University of Michigan). All restriction enzymes, T4 DNA ligase, and Phusion DNA polymerase were purchased from NEB. DNA primers for PCR were purchased from Integrated DNA Technologies. DNA sequencing was performed at the University of Michigan Sequencing Core. Chemical reagents used in enzymatic assays, NADH, DTNB, AcCoA, ATP, phosphoenol pyruvate (PEP), lactic dehydrogenase (LDH) and pyruvate kinase (PK) were purchased from Sigma-Aldrich. The pH was adjusted at room temperature. Spectrophotometric assays were performed on a multimode SpectraMax M5 plate reader using 96-well plates. The Wizard Genomic DNA purification kit (catalog #A1120) used for isolation of H. influenzae ATCC 51907 genomic DNA was purchased from Promega (Madison, WI). The pET16b plasmid containing the CPT gene⁴⁰ used as a template for the cloning of the pCPT-Int-pET19b–pps was provided by Dr. Jacqueline Ellis (University of Leicester). Chemical reactions were monitored by TLC (Merck, Silica gel 60 F₂₅₄) and visualized using a cerium-molybdate stain ((NH₄)₂Ce(NO₃)₆ (5 g), (NH₄)₆Mo₇O₂₄•4H₂O (120 g), H₂SO₄ (80 mL), H₂O (720 mL)). Compounds were purified by SiO₂ flash chromatography (Merck Kieselgel 60). ¹H (including 1D-TOCSY) and ¹³C NMR spectra were recorded on Bruker Avance™ 400 and 500 spectrometers (Figs. S1–S30). Low-resolution electrospray ionization (LRMS (ESI)) mass spectra were recorded on a Waters 3100 mass detector. High-resolution electrospray (ESI-MS) mass spectra were recorded on a Waters Synapt instrument. The purity of the new dimers was determined by reversed-phase high-performance liquid chromatography (RP-HPLC, C18 column, 5 µ, 250×4.6 mm) using a gradient of 1:9 to 9:1 acetonitrile:H₂O (0.1% TFA) over 30 min at 1 mL/min (Table S2 and Figs. S31– S39) and the identity of the compound was confirmed by mass spectrometry. The purity of all of the dimers tested was ≥95%.

Compound 6

Compound 6 was synthesized as previously reported.^{27 1}H NMR (500 MHz, D₂O) δ 5.67 (d, J = 3.6 Hz, 1H, H-1'), 5.01 (d, J = 3.7 Hz, 1H, H-1"), 3.98-3.90 (m, 2H, H-4, H-5"), 3.87 (m, 2H, H-2", H-5'), 3.83-3.72 (m, 3H, H-5, H-4', H-6), 3.72-3.57 (m, 7H, H-4', H-4", H-2', SCH₂CH₂OCH₂CH₂S (4H)), 3.51 (m, 2H, H-1, H-3), 3.42-3.31 (m, 2H, H-3", H-6'), 3.19 (m, 1H, H-6'), 3.02 (ddd, J₁ = 14.1 Hz, J₂ = 6.7 Hz, J₃ = 2.6 Hz, 1H, H-6"), 2.88 (m, 2H, SCH₂CH₂OCH₂CH₂S), 2.82-2.71 (m, 3H, SCH₂CH₂OCH₂CH₂S (2H), H-6"), 2.48 (dt, J₁ = 12.7 Hz, J₂ = 4.3 Hz, 1H, H-2eq), 2.23 (dt, J₁ = 12.1 Hz, J₂ = 4.4 Hz, 1H, H-3eq), 1.95 (q, J = 12.3 Hz, 1H, H-3ax), 1.87 (q, J = 12.6 Hz, 1H, H-2ax) (Fig. S1). ¹³C NMR (125 MHz, D₂O) δ 163.0 (q, J = 35.3 Hz, CF₃CO₂H), 116.3 (q, J = 291.6 Hz, CF₃CO₂H), 100.7, 94.5, 83.8, 77.6, 74.2, 72.6, 71.8, 69.0, 68.0 (2C), 64.4, 54.6, 49.5, 48.4, 47.8, 39.8, 32.7, 31.9, 29.9 (2C), 29.3, 27.8, 23.1 (Fig. S2). HRMS (ESI) m/z calcd for C₂₂H₄₅N₅O₉S₂ [M+H]⁺ 588.2737, found 588.2741.

Compound 8

Bis(2-mercaptoethyl)ether (3.76 mL, 30.35 mmol) was added to a solution of clindamycin hydrochloride (2 g, 4.33 mmol) and Cs₂CO₃ (2.8 g, 8.60 mmol) in DMF (6 mL). The reaction mixture was stirred at 50 °C for 12 h. The reaction progress was monitored by ESI MS by following the disappearance of the starting material ([M+H]⁺, m/z 425.50) and the formation of 8 ([M+H]⁺, m/z 527.58). Upon completion, the reaction mixture was diluted with EtOAc and washed with 0.1 M HCl and brine. The organic layer was dried over MgSO₄ and concentrated under reduced pressure. Purification by flash column chromatography (SiO₂, 0.5:99.5 to 3:97/MeOH:CH₂Cl₂) afforded a mixture of the desired product and traces of a byproduct that could only be separated from the desired product after the acetylation reaction (1.7 g, 74%). The mixture (1.4 g, 2.6 mmol) was dissolved in anhydrous pyridine (4 mL). Ac₂O (2 mL, 21.3 mmol) and a catalytic amount of DMAP were added and the mixture was stirred at rt. The reaction progress was monitored by TLC (9:1/EtOAc:hexane). After 2 h, the reaction mixture was neutralized by adding 1 N HCl and partitioned between H₂O and EtOAc. The organic layer was washed with brine, dried over MgSO₄, and concentrated under reduced pressure. Purification by flash column chromatography (SiO₂, 1:99 to 4:96/MeOH:CH₂Cl₂) afforded 8 (1.60 g, 87%) as a white solid. ¹H NMR (400 MHz, CD₃OD) δ 5.67 (d, J = 5.7 Hz, 1H, H-4), 5.30 (d, J = 3.2 Hz, 1H, H-1), 5.23 (dd, J₁ = 11.0 Hz, J₂ = 5.7 Hz, 1H, H-3), 5.11 (dd, J₁ = 11.0 Hz, J₂ = 3.3 Hz, 1H, H-2), 4.64-4.38 (m, 2H, H-5, H-6), 3.77-3.63 (m, 2H, SCH₂CH₂OCH₂CH₂SAc), 3.61-3.58 (m, 2H, SCH₂CH₂OCH₂CH₂SAc), 3.44 (qt, J₁ = 7.0 Hz, J₂ = 2.1 Hz, 1H, H-7), 3.25 (dd, J₁ = 8.0, J₂ = 6.0 Hz, 1H, H-2'), 3.05-3.16 (m, 2H, SCH₂CH₂OCH₂CH₂SAc), 2.90 (dd, J₁ = 3.8 Hz, J₂ = 1.0 Hz, 1H, H-3'), 2.89-2.81 (m, 1H, SCH₂CH₂OCH₂CH₂SAc), 2.63 (dt, J₁ = 12.6 Hz, J₂ = 6.2 Hz, 1H, SCH₂CH₂OCH₂CH₂SAc), 2.42 (s, 3H, NCH₃), 2.36 (s, 3H, SCH₃), 2.17 (s, 3H, C(=O)CH₃), 2.13 (s, 3H, C(=O)CH₃), 1.95 (s, 3H, C(=O)CH₃), 1.97-2.03 (m, 3H, H-5'a, H-5'b, and H-3'),1.94 (s, 1H, C(=O)CH₃), 1.14 (m, 1H, H-4'), 1.78-1.26 (m, 4H, H-1" (2H), H-2" (2H)), 1.21 (d, J = 7.1 Hz, 3H, H-8), 0.94 (t, J = 6.8 Hz, 3H, H-3") (Fig. S3). ¹³C NMR (100 MHz, CD₃OD) δ 197.1, 177.3, 172.0, 171.6 (2C), 86.0, 72.2, 70.5, 70.1, 69.8, 69.2, 68.9, 68.7, 63.9, 42.1, 41.3, 38.9, 38.1, 36.9, 31.2, 30.5, 29.7, 22.4, 21.0, 20.6 (2C), 14.6 (2C), 13.8 (Fig. S4). LRMS (ESI) m/z cald for C₃₀H₅₁N₂O₁₀S₃ [M+H]⁺ 695.26, found 695.56.

Compound 9

Sodium methoxide (200 µL of 0.5 M solution in MeOH) was added in four portions of 50 µL to a solution of 8 (160 mg, 0.23 mmol) in MeOH (5 mL). The reaction progress was monitored by TLC (1:9/MeOH:CH₂Cl₂). After 3 h, the excess sodium methoxide was quenched with a few drops of glacial AcOH. The solvent was evaporated and the residue purified by flash chromatography (SiO₂, 1:99 to 6:94/MeOH:CH₂Cl₂) to afford 9 (100 mg, 83%) as a white powder. ¹H NMR (400 MHz, CD₃OD) δ 5.28 (d, J = 5.6 Hz, 1H, H-1), 4.46 (dd, J₁ = 9.6 Hz, J₂ = 3.8 Hz, 1H, H-6), 4.17 (dd, J₁ = 9.9 Hz, J₂ = 1.1 Hz, 1H, H-5), 4.13 (dd, J₁ = 10.2 Hz, J₂ = 5.6 Hz, 1H, H-2), 3.8 (dd, J₁ = 3.3 Hz, J₂ = 1.1 Hz, 1H, H-4), 3.69-3.63 (m, 2H, SCH₂CH₂OCH₂CH₂S), 3.62-3.58 (m, 2H, SCH₂CH₂OCH₂CH₂S), 3.56 (dd, J₁ = 3.3 Hz, J₂ = 1.1 Hz, 1H, H-3), 3.45 (qd, J₁ = 7.0 Hz, J₂ = 3.7 Hz, 1H, H-7), 3.25 (dd, J₁ = 8.6 Hz, J₂ = 6.0 Hz, 1H, H-3'), 3.01 (dd, J₁ = 10.5 Hz, J₂ = 5.0 Hz, 1H, H-2'), 2.85 (dt, J₁ = 13.8 Hz, J₂ = 6.3 Hz, 1H, SCH₂CH₂OCH₂CH₂S), 2.78-2.62 (m, 3H, SCH₂CH₂OCH₂CH₂SH), 2.45 (s, 3H, NCH₃), 2.29-2.00 (m, 3H, H-5' (2H), H-3'), 2.15 (s, 3H, SCH₃), 1.87 (ddd, J₁ = 13.0 Hz, J₂ = 10.5 Hz, J₃ = 9.4 Hz, 1H, H-4'), 1.40 (m, 4H, H-1" (2H), H-2" (2H)), 1.31 (d, J = 7.5 Hz, 3H, H-8), 0.94 (m, 3H, H-3") (Fig. S5). ¹³C NMR (100 MHz, CD₃OD) δ 176.9, 88.9, 72.4, 70.6 (2C), 69.4, 69.0, 68.6, 68.1, 62.4, 50.4, 40.6, 39.8, 37.5, 37.3, 35.7, 29.9, 23.4, 21.2, 14.0, 13.2, 12.7 (Fig. S6). HRMS (ESI) m/z calcd for C₂₂H₄₂N₂O₇S₃ [M+H]⁺ 527.2283, found 527.2280.

Compound 10

Sodium methoxide (200 µL of 0.5 M solution in MeOH) was added to a solution of 8 (40 mg, 0.12 mmol) in MeOH (0.15 mL). The reaction progress was monitored by TLC (1:9/MeOH:CH₂Cl₂). When 10 was the major product observed by TLC, a few drops of glacial AcOH were added to the reaction mixture, the solvent was evaporated, and the residue purified by flash chromatography (SiO₂, 1:99 to 10:90/MeOH:CH₂Cl₂) to afford 10 (46 mg, 75%) as a white powder. Purity = 97% (Table S2 and Fig. S31). ¹H NMR (400 MHz, CD₃OD) δ 5.29 (d, J = 5.7 Hz, 2H, H-1), 4.44 (dd, J₁ = 9.5 Hz, J₂ = 3.8 Hz, 2H, H-6), 4.17 (d, J = 9.6 Hz, 1.1 Hz, 2H, H-5), 4.13 (dd, J₁ = 10.2 Hz, J₂ = 5.6 Hz, 2H, H-2), 3.80 (dd, J₁ = 3.3 Hz, J₂ = 1.1 Hz, 2H, H-4), 3.75 (t, J = 6.3 Hz, 4H, SCH₂CH₂OCH₂CH₂S), 3.70-3.65 (m, 4H, SCH₂CH₂OCH₂CH₂S), 3.57 (dd, J₁ = 10.2 Hz, J₂ = 3.4 Hz, 2H, H-3), 3.43 (dd, J₁ = 7.1 Hz, J₂ = 3.9 Hz, 2H, H-7), 3.25 (dd, J₁ = 8.6 Hz, J₂ = 6.0 Hz, 2H, H-5'), 3.00 (dd, J₁ = 10.5 Hz, J₂ = 5.0 Hz, 2H, H-2'), 2.94 (t, J = 6.3 Hz, 4H, SCH₂CH₂OCH₂CH₂S), 2.84 (dt, J₁ = 13.6 Hz, J₂ = 6.5 Hz, 4H, SCH₂CH₂OCH₂CH₂S), 2.72 (dt, J₁ = 13.6 Hz, J₂ = 6.5 Hz, 2H, SCH₂CH₂OCH₂CH₂S), 2.45 (s, 3H, NCH₃), 2.29-2.18 (m, 2H, H-5'), 2.16 (s, 6H, SCH₃), 2.07-1.97 (m, 4H, H-3'), 1.86 (ddd, J₁ = 12.9 Hz, J₂ = 10.5 Hz, J₃ = 9.3 Hz, 2H, H-4'), 1.36 (m, 8H, H-1" (4H), H-2" (4H)), 1.28 (d, J = 7.1 Hz, 6H, H-8), 0.95 (m, 6H, H-3") (Fig. S7). ¹³C NMR (100 MHz, CD₃OD) δ 178.4 (2C), 90.3 (2C), 72.1 (4C), 70.8 (2C), 70.5 (2C), 70.1 (2C), 70.0 (2C), 69.5 (2C), 63.9 (2C), 52.0 (2C), 42.1 (2C), 41.2 (2C), 39.7 (2C), 38.9 (2C), 38.7 (2C), 37.1 (2C), 31.3 (2C), 22.6 (2C), 15.5 (2C), 14.6 (2C), 14.1 (2C) (Fig. S8). HRMS (ESI) m/z calcd for C₄₄H₈₂N₄O₁₂S₆ [M+H]⁺ 1051.4332, found 1051.4320.

Compound 12

This compound was prepared using a modified version of a previously published procedure.³⁸ N,N-Diisopropylethylamine (8 mL, 0.045 mol) was added to 2-amino-1-(4-nitrophenyl)-1,3-propanediol (4 gr, 0.18 mole) dissolved in MeOH (40 mL). The mixture was cooled in an ice bath and chloroacetyl chloride (3 mL, 0.036 mol) was added. The reaction progress was monitored by TLC (1:9/MeOH:CH₂Cl₂). After 5 h, when 12 was observed as the major product, the reaction mixture was partitioned between H₂O and EtOAc. The aqueous layer was separated and extracted with EtOAc. The combined organic layers were washed with brine, dried over MgSO₄, and concentrated under reduced pressure. Purification by flash chromatography (SiO₂, 1:99 to 4:96/MeOH:CH₂Cl₂) afforded 12 (3.6 g, 67%) as a white solid. ¹H NMR (400 MHz, CD₃OD) δ 8.14 (d, J = 8.6 Hz, 2H, H-1), 7.63 (d, J = 8.6 Hz, 1H, H-2), 5.17 (d, J = 2.9 Hz, 1H, H-3), 4.21 (dt, J₁ = 6.7 Hz, J₂ = 3.4 Hz, 1H, H-4), 4.0 (d, J = 14.0 Hz, 2H, H-6), 3.96 (d, J = 14.0 Hz, 2H, H-6) 3.81 (dd, J₁ = 11.1 Hz, J₂ = 6.6 Hz, 1H, H-5), 3.64 (dd, J₁ = 11.1 Hz, J₂ = 6.1 Hz, H-5) (Fig. S9). ¹³C NMR (100 MHz, CD₃OD) δ 169.1, 151.0, 148.1, 128.1 (2C), 124.1 (2C), 71.4, 62.1, 57.8, 43.1 (Fig. S10). LRMS (ESI) m/z cald for C₁₁H₁₂ClN₂O₅ [M-H]⁻ 287.05, found 287.28.

Compound 13

Bis(2-mercaptoethyl)ether (0.85 mL, 6.94 mmol) and a catalytic amount of TBAI were added to a solution of 12 (1.00 g, 3.47 mmol) and K₂CO₃ (1.13 g, 3.47 mmol) in anhydrous THF (7 mL). The reaction mixture was stirred at 50 °C. The reaction progress was monitored by TLC (99:1/EtOAc:MeOH). After 3 h, the reaction mixture was partitioned between H₂O and EtOAc. The organic layer was washed with brine, dried over MgSO₄, and concentrated under reduced pressure. Purification by flash chromatography (SiO₂, 6:4/petroleum ether:EtOAc to 99:1/EtOAc:MeOH) afforded 13 (1.35 g, quant.) as a white solid. ¹H NMR (400 MHz, CD₃OD) δ 8.22 (d, J = 8.7 Hz, 2H, H-1), 7.69 (d, J = 8.7 Hz, 2H, H-2), 5.17 (d, J = 2.6 Hz, 1H, H-3), 4.20 (ddd, J₁ = 7.3 Hz, J₂ = 5.9 Hz, J₃ = 2.7 Hz, 1H, H-4), 3.80 (dd, J₁ = 10.8 Hz, J₂ = 7.3 Hz, 1H, H-5), 3.62 (dd, J₁ = 10.8 Hz, J₂ = 5.9 Hz, 1H, H-5), 3.57 (dt, J₁ = 11.7 Hz, J₂ = 6.3 Hz, 4H, SCH₂CH₂OCH₂CH₂S), 3.23 (d, J = 14.9 Hz, 1H, H-6), 3.18 (d, J = 14.9 Hz, 1H, H-6), 2.66 (t, J = 6.4 Hz, 2H, SCH₂CH₂OCH₂CH₂S), 2.55 (t, J = 6.3 Hz, 2H, SCH₂CH₂OCH₂CH₂S) (Fig. S11). ¹³C NMR (100 MHz, CD₃OD) δ 172.2, 152.0, 148.5, 128.3 (2C), 124.2 (2C), 73.8, 71.6, 71.3, 62.5, 58.0, 36.4, 32.8, 24.7 (Fig. S12). HRMS (ESI) m/z calcd for C₁₅H₂₂N₂O₆S₂ [M+Na]⁺ 413.0817, found 413.0813.

Compound 14

2,2'-Dipyridyldisulfide (140 mg, 0.64 mmol) was added to a solution of compound 13 (125 mg, 0.32 mmol) in MeOH (4 mL), and the reaction mixture was stirred at 23 °C. The reaction progress was monitored by TLC (96:4/EtOAc:MeOH). After 5 h, the solvent was removed under reduced pressure. Purification by flash chromatography (SiO₂, 6:4/petroleum ether:EtOAc to 96:4/EtOAc:MeOH) afforded 14 (125 mg, 78%) as a white solid. ¹H NMR (400 MHz, CD₃OD) δ 8.40 (ddd, J₁ = 4.9 Hz, J₂ = 1.7 Hz, J₃ = 0.9 Hz, 1H, H-1'), 8.19 (d, J = 8.8 Hz, 2H, H-1), 7.93 (dd, J₁ = 8.1 Hz, J₂ = 1.2 Hz, 1H, H-3'), 7.83 (ddd, J₁ = 8.04 Hz, J₂ = 4.9 Hz, J₃ = 1.7, Hz, 1H, H-2'), 7.68 (d, J = 8.8 Hz, 2H, H-2), 7.23 (ddd, J₁ = 7.4 Hz, J₂ = 4.9 Hz, J₃ = 1.1 Hz, 1H, H-4'), 5.18 (d, J = 2.6 Hz, 1H, H-3), 4.21 (ddd, J₁ = 7.4 Hz, J₂ = 5.8 Hz, J₃ = 2.7 Hz, 1H, H-4), 3.80 (dd, J₁ = 10.8 Hz, J₂ = 7.3 Hz, 1H, H-5), 3.65 (m, 3H, H-5, SCH₂CH₂OCH₂CH₂S (2H)), 3.51 (t, J = 6.3 Hz, 2H, SCH₂CH₂OCH₂CH₂S), 3.23 (d, J = 14.9 Hz, 1H, H-6), 3.18 (d, J = 14.9, 1H, H-6), 3.03 (t, J = 6.0 Hz, 2H, SCH₂CH₂OCH₂CH₂S), 2.52 (t, J = 6.4 Hz, 2H, SCH₂CH₂OCH₂CH₂S) (Fig. S13). ¹³C NMR (100 MHz, CD₃OD) δ 172.2, 161.7, 152.0, 150.2, 148.5, 139.2, 128.3 (2C), 124.2 (2C), 122.3, 121.2, 71.6, 71.3, 69.5, 62.5, 58.0, 39.9, 36.5, 32.7 (Fig. S14). LRMS (ESI) m/z calcd for C₂₀H₂₅N₃O₆S₃Cl [M+Cl]⁻ 534.06, found 534.17.

Compound 16

Compound 16 was synthesized as previously reported.²⁷ Purity = 97% (Table S2 and Fig. S32). ¹H NMR (400 MHz, CD₃OD) δ 5.79 (d, J = 3.7 Hz, 2H, H-1'), 5.08 (d, J = 3.7 Hz, 4H, H-1"), 4.13 (t, J = 9.6 Hz, 2H, H-4), 3.99 (m, 4H, H-5', H-5"), 3.94-3.89 (m, 4H, H-2", H-5), 3.82-3.73 (m, 6H, H-5, SCH₂CH₂OCH₂CH₂S (4H)), 3.69 (t, J = 6.4 Hz, 4H, SCH₂CH₂OCH₂CH₂S), 3.58 (m, 10H, H-1, H-3, H-2', H-4', H-4"), 3.46-3.37 (m, 4H, H-6', H-3"), 3.21-3.07 (m, 4H, H-6', H-6"), 2.91 (t, J = 6.3 Hz, 4H, SCH₂CH₂OCH₂CH₂S), 2.87-2.79 (m, 6H, SCH₂CH₂OCH₂CH₂S (4H), H-6"), 2.57 (dt, J₁ = 12.5 Hz, J₂ = 4.3 Hz, 2H, H-2eq), 2.23 (dt, J₁ = 12.0 Hz, J₂ = 4.5 Hz, 2H, H-3eq), 2.19-2.04 (m, 4H, H-3ax, H-3eq) (Fig. S15). ¹³C NMR (125 MHz, D₂O) δ 162.5 (q, J = 36.2 Hz, CF₃CO₂H), 116.0 (q, J = 290.9 Hz, CF₃CO₂H), 100.6 (2C), 94.3 (2C), 83.8 (2C), 77.4 (2C), 74.2 (2C), 72.5 (2C), 70.3 (2C), 69.1 (2C), 67.9 (4C), 64.4 (2C), 54.6 (2C), 49.5 (2C), 48.2 (2C), 47.8 (2C), 39.8 (2C), 37.0 (2C), 32.6 (2C), 31.8 (2C), 29.5 (2C), 29.2 (2C), 27.7 (2C) (Fig. S16). HRMS (ESI) m/z calcd for C₄₄H₈₈N₁₀O₁₈S₄ [M+H]⁺ 1173.5239, found 1173.5245.

Compound 17

A solution of 7 (127 mg, 0.106 mmol) in MeOH (1 mL) was added to a solution of 9 (50 mg, 0.095 mmol) in MeOH (1 mL), and the reaction was stirred at 23 °C. The reaction progress was monitored by TLC (9:1/CH₂Cl₂:MeOH). Upon completion, the solvent was removed under reduced pressure. Purification by flash chromatography (SiO₂, 99:1 to 91:9/CH₂Cl₂:MeOH) afforded 17 (135 mg, 88%) as a white solid. ¹H NMR (400 MHz, CD₃OD) δ 5.27 (d, J = 5.6 Hz, 1H, H-7), 5.13 (bs, 1H, H-1'), 5.06 (d, J = 3.7 Hz, 1H, H-1"), 4.42 (ddd, J₁ = 9.7 Hz, J₂ = 4.0 Hz, J₃ = 1.5 Hz, 1H, H-12), 4.16-4.08 (m, 2H, H-8, H-11), 3.78-3.71 (m, 8H, H-4, H-5', H-2", H-5", SCH₂CH₂OCH₂CH₂S (4H)), 3.68-3.61 (m, 7H, H-5, H-6, H-10, SCH₂CH₂OCH₂CH₂S (4H)), 3.58-3.33 (m, 10H, H-1, H-3, H-2', H-4', H-6', H-3", H-4", H-9, H-13, H-21), 3.23 (dd, J₁ = 8.6 Hz, J₂ = 5.8 Hz, 1H, H-6'), 3.06-3.02 (m, 1H, H-6"), 2.99 (dd, J₁ = 10.5 Hz, J₂ = 4.9 Hz, 1H, H-15), 2.93-2.89 (m, 4H, SCH₂CH₂OCH₂CH₂S), 2.85-2.87 (m, 4H, SCH₂CH₂OCH₂CH₂S), 2.75-2.65 (m, 1H, H-6"), 2.43 (s, 3H, NCH₃), 2.25-1.94 (m, 7H, H-16 (2H), H-21, H-2eq, SCH₃), 1.90-1.80 (m, 1H, H-3eq), 1.68-1.55 (m, 2H, H-2ax, H-3ax), 1.54-1.27 (m, 50H, H-17, 5xCO₂C(CH₃)₃, H-18 (2H), H-19 (2H)), 1.35 (d, J = 7.0 Hz, 3H, H-14), 1.34 (q, J = 2.4 Hz, 3H), 0.92 (m, 3H, H-20). ¹³C NMR (100 MHz, CD₃OD) δ 176.9, 158.0, 157.8, 156.5, 156.3, 148.7, 98.47, 97.99, 88.9, 82.5, 80.9, 79.3, 79.1, 79.0, 78.8, 72.6, 72.1, 70.6, 70.3, 69.4, 69.0, 68.7, 68.5, 68.1, 55.6, 50.5, 50.1, 49.6, 49.6, 40.7, 39.8, 39.8, 38.2, 38.2, 37.5, 37.3, 35.6, 34.4, 33.7, 32.9, 32.1, 29.8 (15C), 21.2, 14.1, 13.2, 12.7. LRMS (ESI) m/z calcd for C₆₉H₁₂₆N₇O₂₅S₅ [M+H]⁺ 1612.73, found 1613.10.

Compound 18

Neat TFA (0.7 mL) was added to compound 17 (30 mg, 0.018 mmol) at rt. After 3 min, the TFA was removed under reduced pressure, and the product was re-dissolved in a minimal volume of H₂O and freeze-dried to afford 18 (32 mg, quant.) as a white foam. Purity = 96% (Table S2 and Fig. S33). ¹H NMR (500 MHz, D₂O) δ 5.67 (d, J = 3.6 Hz, 1H, H-1'), 5.28 (d, J = 5.9 Hz, 1H, H-7), 5.01 (d, J = 3.5 Hz, 1H, H-1"), 4.51 (dd, J₁ = 8.2 Hz, J₂ = 1.9 Hz, 1H, H-12), 4.22 (dt, J₁ = 8.0 Hz, J₂ = 3.6 Hz, 1H, H-11), 4.13 (m, 1H, H-8), 4.08-3.99 (m, 2H, H-4, H-5"), 3.95-3.70 (m, 16H, H-2", H-5', H-5, H-10, H-6, H-4', H-4", H-2', SCH₂CH₂OCH₂CH₂S (8H)), 3.55-2.98 (m, 4H, H-1, H-3, H-9, H-13), 3.40-3.32 (m, 3H, H-3", H-6', H-21), 3.18 (m, 1H, H-6'), 3.02 (m, 1H, H-6"), 2.96 (m, 1H, H-15), 2.87 (s, 3H, NCH₃), 2.86-2.65 (m, 9H, SCH₂CH₂OCH₂CH₂S (8H), H-6"), 2.48 (dt, J₁ = 12.6 Hz, J₂ = 4.2 Hz, 1H, H-2eq), 2.34 (m, 2H, H-21, H-16), 2.22 (m, 2H, H-16, H-3eq), 2.00 (s, 3H, SCH₃), 1.98-1.82 (m, 3H, H-2ax, H-3ax, H-17), 1.38-1.19 (m, 4H, H-18 (2H) H-19 (2H)), 1.14 (d, J = 6.9 Hz, 3H, H-14), 0.81 (m, 3H, H-20) (Fig. S17). ¹³C NMR (125 MHz, D₂O) δ 169.3, 162.8 (q, J = 35.6 Hz, CF₃CO₂H), 116.3 (q, J = 291.7 Hz, CF₃CO₂H), 100.7, 94.5, 88.4, 83.8, 77.6, 74.2, 72.6, 72.3, 70.5, 70.3 (2C), 69.7, 69.2, 68.6, 68.4, 68.0 (2C), 67.7, 64.4, 62.9 (2C), 54.6, 50.4, 49.5, 48.3, 47.8, 41.0, 40.4, 39.8, 37.5, 37.2, 35.8, 33.4, 32.7, 31.8, 30.0, 29.3, 27.8, 20.5, 14.5, 13.5, 13.3, 13.2 (Fig. S18). HRMS (ESI) m/z calcd for C₄₄H₈₅N₇O₁₅S₅ [M+H]⁺ 1112.4786, found 1112.4791.

Compound 19

Cs₂CO₃ (79 mg, 0.24 mmol), compound 12 (70 mg, 0.24 mmol), and catalytic amount of TBAI were added to compound 5 (265 mg, 0.24 mmol) dissolved in anhydrous THF (1 mL). The reaction mixture was stirred at 40 °C. The reaction progress was monitored by TLC (25:5:75/EtOAc:MeOH:CH₂Cl₂). After 16 h, the reaction mixture was partitioned between H₂O and EtOAc. The organic layer was washed with brine, dried over MgSO₄, and concentrated under reduced pressure. Purification by flash chromatography (SiO₂, 90:5:5 to 70:25:5/CH₂Cl₂:EtOAc:MeOH) afforded 19 (235 mg, 73%) as a white solid. ¹H NMR (400 MHz, CD₃OD) δ 8.20 (d, J = 8.6 Hz, 1H, H-7), 7.67 (d, J = 8.8 Hz, 1H, H-8), 5.17 (d, J = 2.6 Hz, 1H, H-9), 5.11 (br s, 1H, H-1'), 5.05 (d, J = 3.7 Hz, 1H, H-1"), 4.19 (ddd, J₁ = 7.8 Hz, J₂ = 6.1 Hz, J₃ = 2.6 Hz, 1H, H-10), 4.11 (m, 1H, H-5"), 3.79 (dd, J₁ = 10.8 Hz, J₂ = 7.3 Hz, 1H, H-11), 3.78-3.33 (m, 17H, H-11, H-2', H-4', H-5', H-6', H-1, H-3, H-4, H-5, H-6, H-3", H-2", H-4", SCH₂CH₂OCH₂CH₂S (4H)), 3.26-3.13 (m, 3H, H-12, H-6'), 3.04 (m, 1H, H-6"), 2.83-2.66 (m, 2H, SCH₂CH₂OCH₂CH₂S), 2.74-2.66 (m, 1H, H-6"), 2.51 (t, J = 6.3 Hz, 2H, SCH₂CH₂OCH₂CH₂S), 2.14 (d, J = 12.9 Hz, 1H, H-2eq), 2.04 (m, 1H, H-3eq), 1.70-1.59 (m, 2H, H-3ax, H-2ax), 1.45 (m, 45H, 5xCO₂C(CH₃)₃). ¹³C NMR (100 MHz, CD₃OD) δ 170.8 (2C), 158.0, 157.8, 156.5, 156.3, 150.6, 147.1, 126.9 (2C), 122.8 (2C), 98.3, 98.0, 82.6, 80.9, 79.0 (4C), 75.6, 72.5, 72.1, 70.7, 70.5, 70.2, 69.9, 65.1, 61.1, 58.1, 56.6, 55.6, 50.1, 49.7, 49.5, 40.6, 35.0, 34.4, 33.6, 32.9, 32.0, 31.2, 27.4 (15C). LRMS (ESI) m/z calcd for C₅₈H₉₈N₇O₂₄S₂ [M+H]⁺ 1340.60, found 1340.67.

Compound 20

Neat TFA (0.7 mL) was added to compound 19 (30 mg, 0.022 mmol) at rt. After 3 min, the TFA was removed under reduced pressure, and the product was re-dissolved in a minimal volume of H₂O and freeze-dried to afford 20 (32 mg, quant.) as a white foam. Purity = 99% (Table S2 and Fig. S34). ¹H NMR (400 MHz, D₂O) δ 8.12 (d, J = 8.7 Hz, 1H, H-7), 7.52 (d, J = 8.5 Hz, 1H, H-8), 5.63 (d, J = 3.6 Hz, 1H, H-1'), 5.07 (d, J = 2.9 Hz, 1H, H-9), 4.97 (d, J = 3.7 Hz, 1H, H-1"), 4.14 (td, J₁ = 6.5 Hz, J₂ = 2.9 Hz, 1H, H-10), 3.92-3.59 (m, 10H, H-2", H-4", H-5", H-11, H-4', H-5', H-4, H-5, H-6), 3.59-3.24 (m, 9H, H-2', H-3", H-6', H-1, H-3, SCH₂CH₂OCH₂CH₂S (4H)), 3.16-3.02 (m, 3H, H-12a, H-12b, H-6'), 2.96 (dd, J₁ = 14.2 Hz, J₂ = 2.5 Hz, 1H, H-6"), 2.66 (m, 3H, SCH₂CH₂OCH₂CH₂S (2H), H-6"), 2.44 (dt, J₁ = 12.6 Hz, J₂ = 4.3 Hz, 1H, H-2eq), 2.34 (dt, J₁ = 13.6 Hz, J₂ = 6.4 Hz, 1H, SCH₂CH₂OCH₂CH₂S), 2.25 (dt, J₁ = 13.6 Hz, J₂ = 6.3 Hz, 1H, SCH₂CH₂OCH₂CH₂S), 2.18 (dt, J₁ = 12.0 Hz, J₂ = 4.3 Hz, 1H, H-3eq), 1.87 (q, J = 12.0 Hz, 1H, H-3ax), 1.82 (q, J = 12.6 Hz, 1H, H-2ax) (Fig. S19). ¹³C NMR (100 MHz, D₂O) δ 172.4, 162.7 (q, J = 35.9 Hz, CF₃CO₂H), 149.4, 147.0, 127.0 (2C), 123.6 (2C), 116.2 (q, J = 291.4 Hz, CF₃CO₂H), 100.6, 94.5, 83.9, 77.6, 74.3, 72.6, 70.8, 69.1, 68.8, 68.0, 64.4, 61.3, 58.1 (2C), 56.5, 54.7, 49.5, 48.3, 47.9, 39.8, 34.9, 32.7, 31.8, 31.0, 29.3, 27.8 (Fig. S20). HRMS (ESI) m/z calcd for C₃₃H₅₇N₇O₁₄S₂ [M+H]⁺ 840.3483, found 840.3841.

Compound 21

A solution of 13 (40 mg, 0.10 mmol) in MeOH (1.5 mL) was added to a solution of 7 (135 mg, 0.11 mmol) in MeOH (1.5 mL), and the reaction was stirred at 23 °C. The reaction progress was monitored by TLC (95:5/CH₂Cl₂:MeOH). Upon completion, the solvent was removed under reduced pressure. Purification by flash chromatography (SiO₂, 99:1 to 94:6/CH₂Cl₂:MeOH) afforded 21 (120 mg, 80%) as a white solid. ¹H NMR (500 MHz, CD₃OD) δ 8.22 (d, J = 8.7 Hz, 1H, H-7), 7.70 (d, J = 8.8 Hz, 1H, H-8), 5.19 (d, J = 2.6 Hz, 1H, H-9), 5.15 (br s, 1H, H-1'), 5.08 (d, J = 3.5 Hz, 1H, H-1"), 4.21 (ddt, J₁ = 7.9 Hz, J₂ = 6.2 Hz, J₃ = 2.7 Hz, 1H, H-10), 4.13 (ddd, J₁ = 9.5 Hz, J₂ = 6.4 Hz, J₃ = 2.7 Hz, 1H, H-5"), 3.84-3.58 (m, 16H, H-11, H-2', H-4', H-5', H-4, H-5, H-6, SCH₂CH₂OCH₂CH₂S (8H)), 3.54-3.43 (m, 5H, H-1, H-3, H-2", H-4", H-6'), 3.23 (d, J = 14.8 Hz, 1H, H-12), 3.17 (d, J = 15.0 Hz, 1H, H-12), 3.07 (dd, J₁ = 14.4 Hz, J₂ = 2.6 Hz, 1H, H-6"), 3.01-2.89 (m, 4H, SCH₂CH₂OCH₂CH₂S), 2.88-2.76 (m, 2H, SCH₂CH₂OCH₂CH₂S), 2.73 (m, 1H, H-6"), 2.54 (t, J = 6.3 Hz, 2H, SCH₂CH₂OCH₂CH₂S), 2.17 (m, 1H, H-2eq), 2.04 (m, 1H, H-3eq), 1.74-1.58 (m, 2H, H-2ax, H-3ax), 1.57-1.37 (m, 45H, 5xCO₂C(CH₃)₃). ¹³C NMR (125 MHz, CD₃OD) δ 170.8 (2C), 156.3(2C), 150.6, 147.1, 126.9 (2C), 122.8 (2C), 97.9 (2C), 79.3 (2C), 79.1, 79.0, 78.8, 75.7, 72.6, 72.1, 70.7, 69.8, 68.7 (2C), 68.7, 65.0, 61.1, 56.6, 55.6, 49.5, 40.6, 38.2, 38.1, 38.0, 35.0, 33.7, 32.9, 32.1, 31.2, 27.4 (15C). LRMS (ESI) m/z calcd for C₆₂H₁₀₅N₇O₂₅S₄Na [M+Na]⁺ 1498.59 found 1498.36.

Compound 22

Neat TFA (0.7 mL) was added to compound 21 (30 mg, 0.02 mmol) at rt. After 3 min, the TFA was removed under reduced pressure, and the product was re-dissolved in a minimal volume of H₂O and freeze-dried to afford 22 (32 mg, quant.) as a white foam. Purity > 99% (Table S2 and Fig. S35). ¹H NMR (400 MHz, CD₃OD) δ 8.22 (d, J = 8.7 Hz, 1H, H-7), 7.69 (d, J = 8.8 Hz, 1H, H-8), 5.79 (d, J = 3.1 Hz, 1H, H-1'), 5.18 (d, J = 2.6 Hz, 1H, H-9), 5.08 (d, J = 3.7 Hz, 1H, H-1"), 4.21 (td, J₁ = 6.1 Hz, J₂ = 3.0 Hz, 1H, H-10), 4.15 (m, 1H, H-4), 3.99 (m, 2H, H-5', H-5"), 3.87 (m, 2H, H-2", H-5), 3.84-3.67 (m, 7H, H-6, H-11, SCH₂CH₂OCH₂CH₂S (4H)), 3.65-3.57 (m, 9H, H-1, H-3, H-4", H-2', H-4', SCH₂CH₂OCH₂CH₂S (4H)), 3.47-3.36 (m, 2H, H-3", H-6'), 3.28-3.06 (m, 4H, H-12, H-6", H-6'), 2.92 (m, 4H, SCH₂CH₂OCH₂CH₂S), 2.83-2.79 (m, 2H, SCH₂CH₂OCH₂CH₂S), 2.82 (m, 1H, SCH₂CH₂OCH₂CH₂S), 2.76 (ddd, J₁ = 14.2 Hz, J₂ = 7.5 Hz, J₃ = 3.4 Hz, 1H, H-6"), 2.60-2.51 (m, 2H, SCH₂CH₂OCH₂CH₂S (1H), H-2eq), 2.23 (dt, J₁ = 11.9 Hz, J₂ = 4.5 Hz, 1H, H-3eq), 1.87 (q, J = 11.9 Hz, 1H, H-3ax), 1.84 (q, J = 12.6 Hz, 1H, H-2ax) (Fig. S21). ¹³C NMR (100 MHz, D₂O) δ 173.1, 163.6 (q, J = 35.9 Hz, CF₃CO₂H), 150.0, 147.7, 127.7 (2C), 124.3 (2C), 117.0 (q, J = 291.6 Hz, CF₃CO₂H), 101.4, 95.2, 84.6, 78.2, 74.9, 73.3, 73.2, 73.0, 71.4, 71.0, 69.9, 69.6, 68.6 (2C), 65.1, 61.9, 57.2, 55.3, 50.2, 49.0, 48.6, 40.8, 40.5, 38.6, 35.7, 33.4, 32.6, 31.7, 30.0, 28.5 (Fig. S22). HRMS (ESI) m/z calcd for C₃₇H₆₅N₇O₁₅S₄ [M+H]⁺ 976.3500, found 976.3506.

Compound 23

Cs₂CO₃ (27 mg, 0.08 mmol), compound 12 (49 mg, 0.170 mmol), and a catalytic amount of TBAI were added to compound 9 (45 mg, 0.085 mmol) dissolved in anhydrous THF (1 mL). The reaction mixture was stirred at 40 °C. The reaction progress was monitored by TLC (2:1:7/EtOAc:MeOH:CH₂Cl₂). After 16 h, the reaction mixture was partitioned between H₂O and EtOAc. The organic layer was washed with brine, dried over MgSO₄, and concentrated under reduced pressure. Purification by flash chromatography (SiO₂, 9:1/CH₂Cl₂:EtOAc to 8:1:1/CH₂Cl₂:EtOAc:MeOH) afforded 23 (47 mg, 71%) as a white solid. Purity = 98% (Table S2 and Fig. S36). ¹H NMR (400 MHz, CD₃OD) δ 8.21 (d, J = 8.8 Hz, 2H, H-9), 7.7 (d, J = 8.9 Hz, 2H, H-10), 5.28 (d, J = 5.7 Hz, 1H, H-1), 5.18 (d, J = 2.7 Hz, 1H, H-11), 4.44 (dd, J₁ = 9.5 Hz, J₂ = 3.9 Hz, 1H, H-6), 4.27-4.16 (m, 2H, H-5, H-12), 4.12 (dd, J₁ = 10.2 Hz, J₂ = 5.6 Hz, 1H, H-2), 3.82-3.74 (m, 2H, H-4, H-13), 3.66-3.54 (m, 6H, H-3, H-13, SCH₂CH₂OCH₂CH₂S (4H)), 3.40 (dd, J₁ = 7.1 Hz, J₂ = 3.9 Hz, 1H, H-7), 3.26-3.11 (m, 3H, H-9, H-14 (2H)), 2.99 (dd, J₁ = 10.5 Hz, J₂ = 5.0 Hz, 1H, H-5'), 2.85 (dt, J₁ = 13.6 Hz, J₂ = 6.5 Hz, 1H, SCH₂CH₂OCH₂CH₂S), 2.74 (dt, J₁ = 13.6 Hz, J₂ = 6.5 Hz, 1H, SCH₂CH₂OCH₂CH₂S), 2.55 (t, J = 6.3 Hz, 2H, SCH₂CH₂OCH₂CH₂S), 2.45 (s, 3H, NCH₃), 2.21-2.15 (m, 1H, H-5'), 2.14 (s, 3H, SCH₃), 2.10-1.96 (m, 2H, H-3' (2H)), 1.86 (m, 1H, H-4'), 1.47-1.28 (m, 4H, H-1" (2H), H-2" (2H)), 1.29 (d, J = 7.1 Hz, 1H, H-8), 0.94 (m, 3H, H-3") (Fig. S23). ¹³C NMR (100 MHz, CD₃OD) δ 178.4, 172.3, 152.1, 148.5, 128.4 (2C), 124.2 (2C), 90.3, 72.1, 71.7, 71.3, 70.8, 70.5, 70.2, 70.1, 70.0, 69.5, 63.8, 62.5, 58.1, 52.0, 42.1, 41.2, 39.7, 39.5, 38.9, 38.7, 37.1, 36.5, 32.7, 31.3, 22.6, 15.5, 14.6, 14.1 (Fig. S24). HRMS (ESI) m/z calcd for C₃₃H₅₄N₄O₁₁S₃ [M+H]⁺ 779.3029, found 779.3031.

Compound 24

A solution of 14 (95 mg, 0.18 mmol) in MeOH (2.5 mL) was added to a solution of 9 (90 mg, 0.18 mmol) in MeOH (2.5 mL), and the reaction was stirred at 23 °C. The reaction progress was monitored by TLC (9:1/CH₂Cl₂:MeOH). Upon completion, the solvent was removed under reduced pressure. Purification by flash chromatography (SiO₂, 99:1 to 88:12/CH₂Cl₂:MeOH) afforded 24 (132 mg, 80%) as a white solid. Purity > 99% (Table S2 and Fig. S37). ¹H NMR (400 MHz, CD₃OD) δ 8.21 (d, J = 8.8 Hz, 2H, H-9), 7.69 (d, J = 8.4 Hz, 2H, H-10), 5.28 (d, J = 5.6 Hz, 1H, H-1), 5.18 (d, J = 2.6 Hz, 1H, H-11), 4.44 (dd, J₁ = 9.6 Hz, J₂ = 3.8 Hz, 1H, H-6), 4.23-4.17 (m, 2H, H-5, H-12), 4.12 (dd, J₁ = 10.2 Hz, J₂ = 5.6 Hz, 1H, H-2), 3.83-3.52 (m, 11H, H-4, H-13 (2H), SCH₂CH₂OCH₂CH₂S (8H)), 3.43 (dp, J₁ = 7.1 Hz, J₂ = 3.6 Hz, 1H, H-7), 3.26-3.10 (m, 3H, H-9, H-14 (2H)), 3.00 (dd, J₁ = 10.6 Hz, J₂ = 4.9 Hz, 1H, H-5'a), 2.93 (m, 4H, SCH₂CH₂OCH₂CH₂S), 2.84 (dt, J₁ = 13.8 Hz, J₂ = 6.4 Hz, 1H, SCH₂CH₂OCH₂CH₂S), 2.72 (dt, J₁ = 13.8 Hz, J₂ = 6.4 Hz, 1H, SCH₂CH₂OCH₂CH₂S), 2.55 (td, J₁ = 6.3 Hz, J₂ = 1.9 Hz, 2H, SCH₂CH₂OCH₂CH₂S), 2.45 (s, 3H, NCH₃), 2.22-2.16 (m, 1H, H-5'b), 2.15 (s, 3H, SCH₃), 2.11-1.96 (m, 2H, H-3' (2H)), 1.86 (m, 1H, H-4'), 1.45-1.31 (m, 4H, H- 1" (2H), H-2" (2H)),1.27 (d, J = 7.1 Hz, H-8), 0.98-0.92 (m, 3H, H-3") (Fig. S25). ¹³C NMR (100 MHz, CD₃OD) δ 178.4, 172.2, 152.0, 148.5, 128.3 (2C), 124.2 (2C), 90.3, 72.0, 71.6, 71.3, 70.8, 70.4, 70.1, 70.0, 69.9, 69.5, 63.8, 62.5, 58.0, 51.9, 42.0, 41.2, 39.6, 39.5, 38.9, 38.7, 37.0, 36.4, 32.6, 31.2, 22.6, 15.4, 14.5, 14.0 (Fig. S26). HRMS (ESI) m/z calcd for C₃₇H₆₂N₄O₁₂S₅ [M+H]⁺ 915.3046, found 915.3052.

Compound 25

Cs₂CO₃ (33 mg, 0.1 mmol), compound 12 (30 mg, 0.1 mmol), and a catalytic amount of TBAI were added to compound 13 (40 mg, 0.1 mmol) dissolved in anhydrous THF (1 mL). The reaction mixture was stirred at 40 °C. The reaction progress was monitored by TLC (96:4/EtOAc:MeOH). After 16 h, the reaction mixture was partitioned between H₂O and EtOAc. The organic layer was washed with brine, dried over MgSO₄, and concentrated under reduced pressure. Purification by flash chromatography (SiO₂, 6:4/petroleum ether: EtOAc to 96:4/EtOAc:MeOH) afforded 25 (45 mg, 68%) as a white solid. Purity = 98% (Table S2 and Fig. S38). ¹H NMR (400 MHz, CD₃OD) δ 8.21 (d, J = 8.7 Hz, 4H, H-1), 7.69 (d, J = 8.5 Hz, 4H, H-2), 5.17 (d, J = 2.7 Hz, 2H, H-3), 4.20 (m, 1H, H-4), 3.80 (dd, J₁ = 10.8 Hz, J₂ = 7.3 Hz, 2H, H-5), 3.62 (dd, J₁ = 10.6 Hz, J₂ = 5.9 Hz, 2H, H-5), 3.54 (t, J = 6.2 Hz, 4H, SCH₂CH₂OCH₂CH₂S), 3.22 (d, J = 14.9 Hz, 2H, H-6), 3.16 (d, J = 14.9 Hz, 2H, H-6), 2.54 (t, J = 6.2 Hz, 4H, SCH₂CH₂OCH₂CH₂S) (Fig. S27). ¹³C NMR (100 MHz CD₃OD) δ 172.3 (2C), 152.0 (2C), 148.6 (2C), 128.4 (4C), 124.2 (4C), 71.4 (4C), 62.5 (2C), 58.1 (2C), 36.4 (2C), 32.7 (2C) (Fig. S28). HRMS (ESI) m/z calcd for C₂₆H₃₄N₄O₁₁S₂ [M+Na]⁺ 665.1563, found 665.1576.

Compound 26

A solution of 14 (40 mg, 0.060 mmol) in MeOH (2 mL) was added to a solution of 13 (35 mg, 0.089) in MeOH (2 mL), and the reaction was stirred at 23 °C. The reaction progress was monitored by TLC (96:4/EtOAc:MeOH). Upon completion, the solvent was removed under reduced pressure. Purification by flash chromatography (SiO₂, 6:4/petroleum ether: EtOAc to 96:4/EtOAc:MeOH) afforded 26 (56 mg, 89%) as a white solid. Purity = 97% (Table S2 and Fig. S39). ¹H NMR (500 MHz, CD₃OD) δ 8.22 (d, J = 8.8 Hz, 4H, H-1), 7.68 (d, J = 8.5 Hz, 4H, H-2), 5.18 (d, J = 2.7 Hz, 2H, H-3), 4.20 (ddd, J₁ = 7.6 Hz, J₂ = 5.8 Hz, J₃ = 2.7 Hz, 2H, H-4), 3.80 (dd, J₁ = 10.8 Hz, J₂ = 7.4 Hz, 2H, H-5'), 3.70 (t, J = 6.4 Hz, 4H, SCH₂CH₂OCH₂CH₂S), 3.65 (dd, J₁ = 10.7 Hz, J₂ = 5.9 Hz, 2H, H-5'), 3.59 (m, 4H, SCH₂CH₂OCH₂CH₂S), 3.23 (d, J = 14.9 Hz, 2H, H-6), 3.16 (d, J = 14.9, 2H, H-6), 2.92 (t, J = 6.3 Hz, 4H, SCH₂CH₂OCH₂CH₂S), 2.54 (t, J = 6.3 Hz, 4H, SCH₂CH₂OCH₂CH₂S) (Fig. S29). ¹³C NMR (125 MHz, CD₃OD) δ 172.3 (2C), 152.1 (2C), 148.5 (2C), 128.3 (4C), 124.2 (4C), 71.7 (2C), 71.3 (2C), 70.1 (2C), 62.5 (2C), 58.1 (2C), 39.5 (2C), 36.4 (2C), 32.7 (2C) (Fig. S30). HRMS (ESI) m/z calcd for C₃₀H₄₂N₄O₁₂S₄ [M+Na]⁺ 801.1580, found 801.1588.

Preparation of the pCNR-Int-pET19b-pps overexpression construct

H. influenzae ATCC 51907 genomic DNA was isolated using the Wizard Genomic DNA purification kit by following the Promega protocol for Gram-negative bacterial samples. To construct the pCNR-Int-pET19b-pps plasmid encoding the chloramphenicol nitroreductase (CNR), PCR was performed using H. influenzae ATCC 51907 genomic DNA as a template, the forward primer (GGAAATCATATGACTCAACTTACTCGT), the reverse primer (TTAATGCTCGAGTTACCCCACCCATTT), and Phusion DNA polymerase. The resulting DNA fragment was inserted into the linearized Int-pET19b-pps by using the corresponding NdeI and XhoI restriction sites (underlined in the primers). The resulting plasmid was transformed into E. coli TOP10 chemically competent cells. The plasmid bearing the CNR gene insert was sequenced and showed perfect alignment with the reported sequence (locus tag HI1278).

Preparation of the pCPT-Int-pET19b-pps overexpression construct

To construct the pCPT-Int-pET19b-pps plasmid encoding the chloramphenicol phosphotransferase (CPT), PCR was performed using pCPT-pET16b as a template, the forward primer (CCCTCTCATATGATCATCCTCAACGGC), the reverse primer (GCAGCCCTCGAGCTACGGGACGACGTG), and Phusion DNA polymerase. The resulting DNA fragment was inserted into the linearized Int-pET19b-pps by using the corresponding NdeI and XhoI restriction sites (underlined in the primers). The resulting plasmid was transformed into E. coli TOP10 chemically competent cells. The plasmid bearing the CPT gene insert was sequenced and showed perfect alignment with the reported sequence (locus tag SVEN_4064).

Overproduction and purification of chloramphenicol-modifying enzymes

The CNR and CPT enzymes (with NHis₁₀ tags) were prepared in a similar manner. To produce a large amount of protein, the pCNR-Int-pET19b-pps and pCPT-Int-pET19b-pps vectors were transformed into E. coli BL21 (DE3) chemically competent cells. For protein overexpression, a 10-mL portion of an overnight culture was used as an inoculum for 1-L of culture of LB supplemented with ampicillin (100 µg/mL). Cells were grown to an absorbance of 0.6 at 600 nm, at which point protein expression was induced with 1 mM IPTG (final concentration). Cultures were incubated overnight at 20 °C. Cells were collected, lysed, and protein was purified by Ni^II-NTA affinity chromatography as we previously reported for the purification of ANT(4').⁴¹ After purification, the proteins were dialyzed in 50 mM Tris-HCl (pH 8.0), 100 mM NaCl, and 10% v/v glycerol and flash frozen in liquid N₂ for storage at −80 °C. CNR was obtained in yields of 3.8 mg/L of culture, and CPT was obtained at 1.4 mg/L of culture (Fig. S40).

Determination of the CNR cofactor

In order to determine the CNR cofactor (FMN or FAD), a sample of the CNR enzyme was boiled (10 min), and the precipitated protein was removed by centrifugation (13,000 rpm, 10 min, rt). The supernatant was split into aliquots (100 µL) into a 96-well plate. Snake venom from Naja naja kaouthia (Sigma V9125) resuspended in 50 mM Tris, pH 8.0 (adding just enough buffer so that the venom was not yellow, ~5 µL) was then added to the supernatant and the fluorescence (ex. 450 nm; em. 520 nm) of FMN was monitored every minute for 30 min. FMN (100 µL, 0.1 mM and 0.5 mM) and FAD (100 µL, 0.1 mM and 0.5 mM), treated with snake venom exactly as was the CNR supernatant, were utilized as controls. An increase in fluorescence was observed for the FAD controls, whereas no change in fluorescence was observed for the FMN controls or the boiled CNR enzyme supernatant, leading to the conclusion that FMN is the cofactor bound to the CNR.

Determination of MIC values of antibiotic compounds

MIC values were determined against a variety of Gram-positive and Gram-negative bacterial strains: B. subtilis 168 (A), B. subtilis 168 with AAC(6')/APH(2")-pRB374 (B), M. smegmatis str. MC2 155 (C), B. cereus ATCC 11778 (D), L. monocytogenes ATCC 19115 (E), S. aureus ATCC 29213 (F), S. aureus NorA (G), E. coli MC1061 (H), B. anthracis 34F2 Sterne strain (I), H. influenzae ATCC 51907 (J), S. epidermidis ATCC 12228 (biofilm negative) (K), and S. pyogenes M12 str. MGAS9429 (L). Strains were tested using a double-dilution of our compounds. All experiments were performed in duplicate or triplicate.

Determination of activity of compounds as inhibitors of bacterial protein synthesis

Protein translation inhibition was quantified in a coupled transcription/translation assay by using E. coli S30 extracts for circular DNA with the pBESTluc plasmid (Promega) according to the manufacturer’s protocol. Translation reactions (10 µL) containing various concentrations of the tested compounds were incubated at 37 °C for 90 min, cooled on ice for 5 min, and diluted with a dilution reagent (25 mM Tris-phosphate buffer, pH 7.8, 2 mM DTT, 2 mM 1,2-diaminocyclohexanetetraacetate, 10% glycerol, 1% Triton X-100, and 1 mg/mL BSA) into 96-well plates. The luminescence was measured immediately after the addition of the luciferase assay reagent (25 µL; Promega), and the light emission was recorded by using a Victor3 plate reader (Perkin-Elmer). The concentrations of half-maximal inhibition (IC₅₀) values were obtained from concentration-response curves fitted to the data of at least two independent experiments by using Grafit 5 software.

Determination of antibiotic-modifying enzyme activity on our antibiotics compounds

Several previously developed assays were employed to visualize the transformation of the compounds by antibiotic-modifying enzymes. All reactions were monitored at 25 °C on a SpectraMax M5 microplate reader and performed in triplicate. All rates were normalized to the parent compound.

Acetylation

The activity of several acetyltransferase enzymes (AAC(6')/APH(2"),⁴² AAC(6')-Ib',⁴³ AAC(3)-IV,⁴² AAC(2')-Ic,⁴⁴ Eis,⁴⁴ and CAT_I⁴⁵) was monitored at 412 nm using Ellman’s method, coupling the release of CoASH with DTNB. Due to the reactivity of the thiol moiety in the acetylation assay, compounds 6, 9, and 13 bearing free thiols were not tested in this assay. Briefly, reactions (200 µL) containing antibiotic compounds (100 µM), AcCoA (500 µM for Eis and 150 µM for other acetyltransferases) were incubated with acetyltransferase enzyme (0.125 µM AAC(3)-IV and AAC(2')-Ic, 0.5 µM for other acetyltransferases) in the presence of DTNB (2 mM) and the appropriate buffer (50 mM MES pH 6.6 for AAC(6')/APH(2") and AAC(3)-IV; 50 mM Tris pH 7.5 for AAC(6')-Ib'; 100 mM sodium phosphate pH 7.4 for AAC(2')-Ic; and 50 mM Tris pH 8.0 for Eis and CAT_I). Measurements were taken every 30 sec for 1 h. Initial rates of the reactions were calculated using data from the first 2–5 min of the reaction.

Nucleotidylation

The nucleotidylation activity of ANT(4') from S. aureus was monitored through the formation of a complex of molybdate, malachite green, and the inorganic phosphate (P_i) generated by inorganic pyrophosphatase (Sigma-Aldrich catalog #I1643) cleavage of the released pyrophosphate (PP_i) during the nucleotidylation reaction.⁴¹ To analyze the activity of ANT(4') on compounds 1, 6, 16, 18, 20, and 22, reactions (160 µL) containing Tris-HCl (50 mM, pH 7.5), MgCl₂ (10 mM), KCl (40 mM), inorganic pyrophosphatase (0.2 U/mL), antibiotic compound (100 µM), and ATP (0.5 mM) were performed at 25 °C. The reactions were initiated by the addition of the enzyme (1 µM), incubated for 20, 40, 60, 120, and 140 s, and quenched by the addition of the malachite green reagent (40 µL). After 15 min of color development, the liberated P_i concentration was measured at 600 nm. The initial rates were determined using the first 60 sec of the reaction.

Phosphorylation

The phosphorylation activity of CPT was monitored at 340 nm through the consumption of NADH in an enzyme-coupled response to the production of ADP.⁴⁶ Reactions (250 µL total volume) contained 100 µM antibiotic compound, 50 mM Tris-HCl (pH 8.0), 10 mM MgCl₂, 40 mM KCl, 0.5 mg/mL NADH, 2.5 mM PEP, 2 mM ATP, and 5 µL PK/LDH, and were initiated by addition of 1 µM CPT. The progress of the reaction was monitored by taking readings every 30 sec for 1 h. The initial rates were determined using the first 5 min of the reaction.

Reduction

The reduction activity of CNR was monitored at 340 nm through the consumption of NADH used to reduce the FMN in the active site of CNR. Reactions (200 µL total volume) contained 100 µM antibiotic compound, 50 mM Tris-HCl (pH 7.5) and 0.5 mg/mL NADH and were initiated by addition of 0.5 µM CNR. Reaction progress was monitored every 30 sec for 1 h. The initial rates were determined using the first 5 min of the reaction.

ABBREVIATIONS

AAC

aminoglycoside acetyltransferase

AG

aminoglycoside

AME

aminoglycoside-modifying enzyme

ANT

aminoglycoside nucleotidyltransferase

APH

aminoglycoside phosphotransferase

CAT

chloramphenicol acetyltransferase

CNR

chloramphenicol nitroreductase

CPT

chloramphenicol phosphotransferase